Among the techniques widely used for thin-film deposition is Chemical Vapor Deposition (CVD) that uses chemically reactive molecules that react in a reaction chamber to deposit a desired film on a substrate. Molecular precursors useful for CVD applications comprise elemental (atomic) constituents of the film to be deposited and typically also include additional elements. CVD precursors are volatile molecules that are delivered, in a gaseous phase, to a chamber in order to react at the substrate, forming the thin film thereon. The chemical reaction deposits a thin film with a desired film thickness.
Common to most CVD techniques is the need for application of a well-controlled flux of one or more molecular precursors into the CVD reactor. A substrate is kept at a well-controlled temperature under controlled pressure conditions to promote chemical reaction between these molecular precursors, concurrent with efficient removal of byproducts. Obtaining optimum CVD performance requires the ability to achieve and sustain steady-state conditions of gas flow, temperature, and pressure throughout the process, and the ability to minimize or eliminate transients.
Especially in the field of semiconductor, integrated circuit, and other electronic devices, there is a demand for thin films, especially higher quality, denser films, with superior conformal coating properties, beyond the achievable limits of conventional CVD techniques, especially thin films that can be manufactured at lower temperatures.
Atomic layer deposition (“ALD”) is an alternative film deposition technology that can provide improved thickness resolution and conformal capabilities compared to its CVD predecessor. The ALD process segments the conventional thin-film deposition process of conventional CVD into single atomic-layer deposition steps. Advantageously, ALD steps are self-terminating and can deposit one atomic layer when conducted up to or beyond self-termination exposure times. An atomic layer typically ranges from about 0.1 to about 0.5 molecular monolayers, with typical dimensions on the order of no more than a few Ångstroms (Å). In ALD, deposition of an atomic layer is the outcome of a chemical reaction between a reactive molecular precursor and the substrate. In each separate ALD reaction-deposition step, the net reaction deposits the desired atomic layer and substantially eliminates “extra” atoms originally included in the molecular precursor. In its most pure form, ALD involves the adsorption and reaction of each of the precursors in the absence of the other precursor or precursors of the reaction. In practice, in any system it is difficult to avoid some direct reaction of the different precursors leading to a small amount of chemical vapor deposition reaction. The goal of any system claiming to perform ALD is to obtain device performance and attributes commensurate with an ALD system while recognizing that a small amount of CVD reaction can be tolerated.
In ALD applications, typically two molecular precursors are introduced into the ALD reactor in separate stages. For example, a metal precursor molecule, MLx, comprises a metal element, M that is bonded to an atomic or molecular ligand, L. For example, M could be, but would not be restricted to, Al, W, Ta, Si, Zn, etc. The metal precursor reacts with the substrate when the substrate surface is prepared to react directly with the molecular precursor. For example, the substrate surface typically is prepared to include hydrogen-containing ligands, AH or the like, that are reactive with the metal precursor. Sulfur (S), oxygen (O), and Nitrogen (N) are some typical A species. The gaseous metal precursor molecule effectively reacts with all of the ligands on the substrate surface, resulting in deposition of a single atomic layer of the metal:substrate-AH+MLx→substrate-AMLx−1+HL  (1)where HL is a reaction by-product. During the reaction, the initial surface ligands, AH, are consumed, and the surface becomes covered with L ligands, which cannot further react with metal precursor MLx. Therefore, the reaction self-terminates when all of the initial AH ligands on the surface are replaced with AMLx−1 species. The reaction stage is typically followed by an inert-gas purge stage that eliminates the excess metal precursor from the chamber prior to the separate introduction of a second reactant gaseous precursor material.
The second molecular precursor then is used to restore the surface reactivity of the substrate towards the metal precursor. This is done, for example, by removing the L ligands and redepositing AH ligands. In this case, the second precursor typically comprises the desired (usually nonmetallic) element A (i.e., O, N, S), and hydrogen (i.e., H2O, NH3, H2S). The next reaction is as follows:substrate-A-ML+AHγ→substrate-A-M-AH+HL  (2)This converts the surface back to its AH-covered state. (Here, for the sake of simplicity, the chemical reactions are not balanced.) The desired additional element, A, is incorporated into the film and the undesired ligands, L, are eliminated as volatile by-products. Once again, the reaction consumes the reactive sites (this time, the L terminated sites) and self-terminates when the reactive sites on the substrate are entirely depleted. The second molecular precursor then is removed from the deposition chamber by flowing inert purge-gas in a second purge stage.
In summary, then, the basic ALD process requires alternating, in sequence, the flux of chemicals to the substrate. The representative ALD process, as discussed above, is a cycle having four different operational stages:    1. MLx reaction;    2. MLx purge;    3. AHy reaction; and    4. AHy purge, and then back to stage 1.
ALD has been most typically utilized for the deposition of inorganic compounds where metal precursors have been halides, alkoxides, diketonate chelates, or organometallic compounds. The second precursor has been typically an oxygen, nitrogen or sulfur source, when oxides, nitrides, or sulfides are deposited, respectively. Although it is relatively less common, the deposition of organic compounds or organic/inorganic hybrid layers by ALD is also known. In these cases, it is possible to still have an alternating sequence of self-limiting reactions, except that the limiting layer produced by such a process may be a layer of molecules as opposed to atoms. Accordingly, such techniques may also be referred to as molecular layer deposition (MLD), although the basic concepts and deposition equipment are similar to ALD processes and equipment and so “ALD” is also used herein to refer to such molecular layer deposition. An example of atomic layer (molecular layer deposition) of organic films can be found in “Atomic Layer Deposition of Polyimide Thin Films,” by Matti Putkonen, et. al. in the Journal of Materials Chemistry. 
The repeated sequence of alternating surface reactions and precursor-removal that restores the substrate surface to its initial reactive state, with intervening purge operations, is a typical ALD deposition cycle. A key feature of ALD operation is the restoration of the substrate to its initial surface chemistry condition. Using this repeated set of steps, a film can be layered onto the substrate in equal metered layers that are all alike in chemical kinetics, deposition per cycle, composition, and thickness.
ALD can be used as a fabrication step for forming a number of types of thin film electronic devices, including semiconductor devices and supporting electronic components such as resistors and capacitors, insulators, bus lines, and other conductive structures. ALD is particularly suited for forming thin layers of metal oxides in the components of electronic devices. General classes of functional materials that can be deposited with ALD include conductors, dielectrics or insulators, and semiconductors.
Conductors can be any useful conductive material. For example, the conductors may comprise transparent materials such as indium-tin oxide (ITO), doped zinc oxide ZnO, SnO2, or In2O3. The thickness of the conductor may vary, and according to particular examples it can range from about 50 to about 1000 nm.
Examples of useful semiconducting materials are compound semiconductors such as gallium arsenide, gallium nitride, cadmium sulfide, intrinsic zinc oxide, and zinc sulfide.
A dielectric material electrically insulates various portions of a patterned circuit. A dielectric layer may also be referred to as an insulator or insulating layer. Specific examples of materials useful as dielectrics include strontiates, tantalates, titanates, zirconates, aluminum oxides, silicon oxides, tantalum oxides, hafnium oxides, titanium oxides, zinc selenide, and zinc sulfide. In addition, alloys, combinations, and multilayers of these examples can be used as dielectrics. Of these materials, aluminum oxides are preferred.
A dielectric structure layer may comprise two or more layers having different dielectric constants. Such insulators are discussed in U.S. Pat. No. 5,981,970 hereby incorporated by reference and copending US Patent Publication No. 2006/0214154, hereby incorporated by reference. Dielectric materials typically exhibit a band-gap of greater than about 5 eV. The thickness of a useful dielectric layer may vary, and according to particular examples it can range from about 10 to about 300 nm.
A number of device structures can be made with the functional layers described above. A resistor can be fabricated by selecting a conducting material with moderate to poor conductivity. A capacitor can be made by placing a dielectric between two conductors. A diode can be made by placing two semiconductors of complementary carrier type between two conducting electrodes. There may also be disposed between the semiconductors of complementary carrier type a semiconductor region that is intrinsic, indicating that that region has low numbers of free charge carriers. A diode may also be constructed by placing a single semiconductor between two conductors, where one of the conductor/semiconductors interfaces produces a Schottky barrier that impedes current flow strongly in one direction. A transistor may be made by placing upon a conductor (the gate) an insulating layer followed by a semiconducting layer. If two or more additional conductor electrodes (source and drain) are placed spaced apart in contact with the top semiconductor layer, a transistor can be formed. Any of the above devices can be created in various configurations as long as the necessary interfaces are created.
In typical applications of a thin film transistor, the need is for a switch that can control the flow of current through the device. As such, it is desired that when the switch is turned on, a high current can flow through the device. The extent of current flow is related to the semiconductor charge carrier mobility. When the device is turned off, it is desirable that the current flow be very small. This is related to the charge carrier concentration. Furthermore, it is generally preferable that visible light have little or no influence on thin-film transistor response. In order for this to be true, the semiconductor band gap must be sufficiently large (>3 eV) so that exposure to visible light does not cause an inter-band transition. A material that is capable of yielding a high mobility, low carrier concentration, and high band gap is ZnO. Furthermore, for high-volume manufacture onto a moving web, it is highly desirable that chemistries used in the process be both inexpensive and of low toxicity, which can be satisfied by the use of ZnO and the majority of its precursors.
Self-saturating surface reactions make ALD relatively insensitive to transport non-uniformities, which might otherwise impair surface uniformity, due to engineering tolerances and the limitations of the flow system or related to surface topography (that is, deposition into three dimensional, high aspect ratio structures). As a general rule, a non-uniform flux of chemicals in a reactive process generally results in different completion times over different portions of the surface area. However, with ALD, each of the reactions is allowed to complete on the entire substrate surface. Thus, differences in completion kinetics impose no penalty on uniformity. This is because the areas that are first to complete the reaction self-terminate the reaction; other areas are able to continue until the full treated surface undergoes the intended reaction.
Typically, an ALD process deposits about 0.1-0.2 nm of a film in a single ALD cycle (with one cycle having numbered steps 1 through 4 as listed earlier). A useful and economically feasible cycle time must be achieved in order to provide a uniform film thickness in a range of from about 3 nm to 30 nm for many or most semiconductor applications, and even thicker films for other applications. According to industry throughput standards, substrates are preferably processed within 2 minutes to 3 minutes, which means that ALD cycle times must be in a range from about 0.6 seconds to about 6 seconds.
ALD offers considerable promise for providing a controlled level of highly uniform thin film deposition. However, in spite of its inherent technical capabilities and advantages, a number of technical hurdles still remain. One important consideration relates to the number of cycles needed. Because of its repeated reactant and purge cycles, effective use of ALD has required an apparatus that is capable of abruptly changing the flux of chemicals from MLx to AHy, along with quickly performing purge cycles. Conventional ALD systems are designed to rapidly cycle the different gaseous substances onto the substrate in the needed sequence. However, it is difficult to obtain a reliable scheme for introducing the needed series of gaseous formulations into a chamber at the needed speeds and without some unwanted mixing. Furthermore, an ALD apparatus must be able to execute this rapid sequencing efficiently and reliably for many cycles in order to allow cost-effective coating of many substrates.
In an effort to minimize the time that an ALD reaction needs to reach self-termination, at any given reaction temperature, one approach has been to maximize the flux of chemicals flowing into the ALD reactor, using so-called “pulsing” systems. In order to maximize the flux of chemicals into the ALD reactor, it is advantageous to introduce the molecular precursors into the ALD reactor with minimum dilution of inert gas and at high pressures. However, these measures work against the need to achieve short cycle times and the rapid removal of these molecular precursors from the ALD reactor. Rapid removal in turn dictates that gas residence time in the ALD reactor be minimized. Gas residence times, τ, are proportional to the volume of the reactor, V, the pressure, P, in the ALD reactor, and the inverse of the flow, Q, that is:τ=VP/Q  (3)
In a typical ALD chamber the volume (V) and pressure (P) are dictated independently by the mechanical and pumping constraints, leading to difficulty in precisely controlling the residence time to low values. Accordingly, lowering pressure (P) in the ALD reactor facilitates low gas residence times and increases the speed of removal (purge) of chemical precursor from the ALD reactor. In contrast, minimizing the ALD reaction time requires maximizing the flux of chemical precursors into the ALD reactor through the use of a high pressure within the ALD reactor. In addition, both gas residence time and chemical usage efficiency are inversely proportional to the flow. Thus, while lowering flow can increase efficiency, it also increases gas residence time.
Existing ALD approaches have been compromised with the trade-off between the need to shorten reaction times with improved chemical utilization efficiency, and, on the other hand, the need to minimize purge-gas residence and chemical removal times. One approach to overcome the inherent limitations of “pulsed” delivery of gaseous material is to provide each reactant gas continuously and to move the substrate through a region containing each gas in succession. In these systems, some mechanism must be employed to confine a particular gas to a spatial region in order that the substrate can sample all of the gases during its movement, but the individual mutually reactive gases cannot mix causing undesirable CVD deposition. Such systems can be referred to as spatially confined ALD systems. For example, U.S. Pat. No. 6,821,563 entitled “GAS DISTRIBUTION SYSTEM FOR CYCLICAL LAYER DEPOSITION” to Yudovsky, describes a processing chamber, under vacuum, having separate gas ports for precursor and purge gases, alternating with vacuum pump ports between each gas port. Each gas port directs its stream of gas vertically downward toward a substrate. The separate gas flows are separated by walls or partitions, with vacuum pumps for evacuating gas on both sides of each gas stream. A lower portion of each partition extends close to the substrate, for example, about 0.5 mm or greater from the substrate surface. In this manner, the lower portions of the partitions are separated from the substrate surface by a distance sufficient to allow the gas streams to flow around the lower portions toward the vacuum ports after the gas streams react with the substrate surface.
A rotary turntable or other transport device is provided for holding one or more substrate wafers. With this arrangement, the substrate is shuttled beneath the different gas streams, effecting ALD deposition thereby. In one embodiment, the substrate is moved in a linear path through a chamber, in which the substrate is passed back and forth a number of times.
Another approach using continuous gas flow is shown in U.S. Pat. No. 4,413,022 entitled “METHOD FOR PERFORMING GROWTH OF COMPOUND THIN FILMS” to Suntola et al. A gas flow array is provided with alternating source gas openings, carrier gas openings, and vacuum exhaust openings. Reciprocating motion of the substrate over the array effects ALD deposition, again, without the need for pulsed operation. In the embodiment of FIGS. 13 and 14, in particular, sequential interactions between a substrate surface and reactive vapors are made by a reciprocating motion of the substrate over a fixed array of source openings. Diffusion barriers are formed by having a carrier gas opening between exhaust openings. Suntola et al. state that operation with such an embodiment is possible even at atmospheric pressure, although little or no details of the process, or examples, are provided.
While systems such as those described in the '563 Yudovsky and '022 Suntola et al. patents may avoid some of the difficulties inherent to pulsed gas approaches, these systems have other drawbacks. Neither the gas flow delivery unit of the '563 Yudovsky patent nor the gas flow array of the '022 Suntola et al. patent can be used in closer proximity to the substrate than about 0.5 mm. Neither of the gas flow delivery apparatus disclosed in the '563 Yudovsky and '022 Suntola et al. patents are arranged for possible use with a moving web surface, such as could be used as a flexible substrate for forming electronic circuits, light sensors, or displays, for example. The complex arrangements of both the gas flow delivery unit of the '563 Yudovsky patent and the gas flow array of the '022 Suntola et al. patent, each providing both gas flow and vacuum, make these solutions difficult to implement, costly to scale, and limit their potential usability to deposition applications onto a moving substrate of limited dimensions. Moreover, it would be very difficult to maintain a uniform vacuum at different points in an array and to maintain synchronous gas flow and vacuum at complementary pressures, thus compromising the uniformity of gas flux that is provided to the substrate surface.
US Patent Publication No. 2005/0084610 to Selitser discloses an atmospheric pressure atomic layer chemical vapor deposition process. Selitser state that extraordinary increases in reaction rates are obtained by changing the operating pressure to atmospheric pressure, which will involve orders of magnitude increase in the concentration of reactants, with consequent enhancement of surface reactant rates. The embodiments of Selitser involve separate chambers for each stage of the process, although FIG. 10 in 2005/0084610 shows an embodiment in which chamber walls are removed. A series of separated injectors are spaced around a rotating circular substrate holder track. Each injector incorporates independently operated reactant, purging, and exhaust gas manifolds and controls and acts as one complete mono-layer deposition and reactant purge cycle for each substrate as is passes there under in the process. Little or no specific details of the gas injectors or manifolds are described by Selitser, although it is stated that spacing of the injectors is selected so that cross-contamination from adjacent injectors is prevented by purging gas flows and exhaust manifolds incorporated in each injector.
Another approach for spatially confining gases in an ALD processing device is described in the above-cited US Publication No. 2007/0228470 which discloses a transverse flow ALD device. In such a device, various gases are directed parallel to each other and thus limit any gas intermixing by limiting the degree of countercurrent flow.
An efficient method for allowing for gas isolation is the floating-head ALD device of the above-cited U.S. patent application Ser. No. 11/620,738. In this device, the pressure of flowing reactive and purge gases is used as a means to separate the coating head from the substrate. Due to the relatively large pressures that can be generated in such a system, gases are forced to travel in well-defined paths and thus eliminate undesired gas intermixing.
Commonly assigned U.S. Pat. Nos. 5,838,345 to Kerr et al., 5,997,119 to Kerr, and 6,677,975 to Kerr relate to apparatus for maintaining the positional relationship of a print head or an imaging subsystem, all three patents hereby incorporated by reference in their entirety. For example, U.S. Pat. No. 5,838,345, relates to an apparatus for maintaining the positional relationship of a print head in a lathe bed scanner, the apparatus comprising a first bar for supporting the print head of the lathe bed scanner. A magnet is positioned adjacent the first bar for magnetically attracting the first bar which, in turn, maintains the positional relationship of the bar and, consequently, the print head.